The Intensive Care Manual, MICHAEL J. APOSTOLAKOS

2 / Intravascular Access and Hemodynamic Monitoring 35 or right atrium in a supine patient; this is where intravascular pressure reaches zero and is independent of body habitus. Although changes in posture can be ex- pected to affect the reference pressure at the phlebostatic axis by less than 1 mm Hg, the CVP is a less accurate indicator of filling pressures when it is measured in the lateral or upright position, because of venous pooling. The CVP is most often used as an approximation of preload and reflects a balance between venous return and right-sided cardiac output. Under normal conditions, the right side of the heart is composed of a thin wall of myocardium and is more compliant than the more muscular left side of the heart. Since CVP measures intravascular pres- sure and not transmural pressure, which is the actual determinant of ventricular preload, its validity as an index of preload is influenced by pulmonary variables, such as intrathoracic pressure, and by cardiac variables, such as cardiac compli- ance. Central filling pressures, such as the CVP, pulmonary artery wedge pressure (PAWP), and pulmonary capillary wedge pressure (PCWP), are measured at end-expiration, when the relative intrathoracic pressure is zero (i.e., it equals atmospheric pressure), and therefore intravascular pressure equals transmural pressure. High levels of positive pressure ventilation, which affect the CVP, should never be discontinued to determine a “more accurate” CVP. In instances where the CVP is thought to be falsely elevated by intrathoracic pressure, an al- ternative form of preload assessment should be considered or esophageal manometry should be used to estimate transthoracic pressure. The transthoracic pressure can then be subtracted from the CVP to provide a better estimate of preload. Graphic depiction of the CVP (also the PCWP and left atrial pressure) wave- forms consists of three positive wave deflections (a, c, and v) and two descents (x and y) (Figure 2–3). The a wave is the increase in venous pressure that is gen- erated by atrial contraction. The c wave occurs when the atrioventricular valve (tricuspid or mitral) is displaced into the atrium during isovolumetric ventricu- lar contraction. The v wave reflects the increase in atrial pressure that occurs as venous return begins to fill the atrium during isovolumetric relaxation, while the atrioventricular valves are still closed. The x descent corresponds to ventricular ejection, as the emptying ventricle draws down on the floor of the atrium and de- creases the CVP. The y descent occurs as the atrioventricular valve opens and blood enters the ventricle during ventricular diastole. The importance of these waveforms lies in their ability to reflect on patho- physiologic processes. Absence of the a wave occurs in atrial fibrillation, in which case the x descent may also be absent. Amplified, or “cannon,” a waves occur in the presence of stenosis of the atrioventricular (mitral) wave. Both the x and y descents are exaggerated in the presence of constrictive pericarditis, whereas car- diac tamponade magnifies the x descent and abolishes the y descent. In the presence of atrioventricular valve incompetency, free transduction of ventricular pressure during ventricular contraction generates large “cannon” V waves that are pathognomonic for regurgitant flow, especially mitral regurgita-

36 The Intensive Care Manual FIGURE 2–3 The CVP wave form as it relates to the electrocardiogram (see text). tion. In the case of the CVP, pulmonary hypertension increases right ventricular afterload, decreases right ventricular compliance, and accentuates the v wave- form depicted on the monitor. PULMONARY ARTERY CATHETER The pulmonary artery catheter (PAC), or Swan-Ganz catheter, provides a more accurate measure of left ventricular preload; however, it also is subject to opera- tor bias and misinterpretation.7,8 The pulmonary artery catheter was originally introduced in 1970 by Swan and Ganz, whose names are still attached to the catheter today. The importance of basing clinical interventions on judicious in- terpretation of the data obtained from the PAC cannot be overemphasized.9,10 Although the discipline of critical care medicine is to a large extent rooted in use of the PAC, recent suggestions that the use of the PAC in clinical medicine is as- sociated with increased morbidity and mortality may reflect, in large part, deci- sions made on the basis of inaccurate data alone, without sufficient consideration of the underlying physiologic principles.

2 / Intravascular Access and Hemodynamic Monitoring 37 Pulmonary Artery Catheter Placement The PAC is passed into the central venous circulation through an introducer catheter, or cordis, and is then passed sequentially through the great veins, the right atrium, right ventricle, pulmonary outflow tract, and into a pulmonary artery. A 1.5-mL silastic balloon allows catheter placement to be flow-directed, because the balloon tip of the catheter, inflated during catheter passage, facilitates placement in the pulmonary outflow tract. Fluoroscopic assistance during place- ment may be indicated if a transvenous pacemaker has been placed recently, se- lective pulmonary artery catheterization is necessary, or anatomic abnormalities, such as Eisenmenger’s complex, exist. Catheterization of the right side of the heart carries the additional risks of ar- rhythmias, intravascular coiling and knotting, and vascular perforation. Con- tinuous waveform analysis of the pressures transduced at the PAC tip allows subjective assessment of the location of the catheter tip. Progression of the catheter tip through the right side of the heart must be monitored by transduced waveform analysis (Figure 2–4). Since the placement of the catheter is flow-directed, advancement of the catheter incrementally with each heartbeat facilitates appropriate passage. Catheter ad- vancement without concomitant waveform progression strongly correlates with placement in the IVC or coiling within the right heart chambers. When the catheter FIGURE 2–4 Waveform analysis during pulmonary artery catheterization. The pressures transduced sequentially include the superior vena cava (SVC) and right atrium (RA) which are typical CVP readings. Entry into the right ventricle (RV) is marked by a rise in the systolic component. The characteristic rise in diastolic pressure signals entry into the pulmonary artery (PA). With the balloon inflated, ‘wedge’ positioning of the balloon tip is signaled by a flattening of the PA waveform. Deflation of the pulmonary artery balloon in the wedge posi- tion should be accompanied by a return to the PA waveform, as blood flow resumes past the catheter tip.

38 The Intensive Care Manual reaches the distal pulmonary artery, the diastolic pressure characteristically rises. Further advancement of the catheter causes the waveform to flatten and signifies that the “wedge” position has been reached; at this point, the balloon occludes the flow of blood past the catheter tip. “Pseudo-wedging” may occur if the catheter is caught underneath the pulmonary valve or trabeculae or between papillary muscles. In this case, waveform flattening occurs prior to pulmonary artery waveform iden- tification. Deflation and reinflation of the balloon is critical, since distal migration of the catheter tip occurs frequently. If slow inflation of the balloon results in a con- tinued rise of the transduced pressure to high levels, the catheter tip is either “over- wedged” in the pulmonary capillary, which carries a high risk of pulmonary artery rupture, or the balloon has herniated past the tip of the catheter, where pressure transduction occurs. Suspicion of “overwedging” requires that inflation attempts be immediately abandoned, the catheter withdrawn a short distance into the pul- monary artery, and the wedge position re-ascertained by slow re-advancement of the catheter. The Physiology and Analysis of Pulmonary Catheter Data PULMONARY CAPILLARY WEDGE PRESSURE AND CARDIAC FUNCTION The pressure determined from this “wedge” waveform at end-expiration is the PCWP, and may be used as an index of left atrial pressure, and by further extrap- olation, the left ventricular end-diastolic pressure. However, true left ventricular preload is actually ventricular wall tension caused by ventricular end-diastolic filling volume, which stretches myocardial sarcomeres. The relationship of ventricular performance to isometric preload is Starling’s law of the heart, and the resulting graphic depiction of this relationship is re- ferred to as a Starling curve (Figure 2–5). Interpretation of the data obtained from use of the PAC is based on the Starling curve and is the foundation for car- diovascular critical care. A fundamental concept of cardiac physiology is that optimal preload develops a tension on the muscle, which causes the overlap of actin and myosin in the myocyte to approximate 2.2 µm. Otherwise, and more practically stated, optimal preload is that precontraction load, or tension, that optimizes ventricular perfor- mance. In graphic terms, optimal preload is the volume that produces ventricular distension nearing the apex of the Starling curve, maximally increasing cardiac contractile function. Overdistention of the sarcomeres beyond 3.0 µm causes a decrease in contractile performance and a negative slope in the Starling curve. Since the pulmonary artery catheter measures pressure, the corresponding vol- ume preload can be inferred only if compliance remains constant during the pe- riod of measurement (i.e., compliance = volume/ pressure). The pulmonary artery catheter has as its greatest utility the ability to depict a mathematical and graphic relationship between cardiac filling pressure and car- diac performance. Data is most reliable when it is directly measured, and mathe- matical manipulation sequentially introduces error. The use of indexed values,

2 / Intravascular Access and Hemodynamic Monitoring 39 FIGURE 2–5 Curvilinear depiction of Starlings law of the heart: The Starling curve relates preload (CVP, PCWP, LVEDP, or LVEDV) to cardiac function (EF, SV, CO) and forms the basis of cardiovascular critical care since both dependent and independent variables can be tracked using a PAC. The incremental increases in preload (a, b, c) are accompanied by corre- sponding increases in cardiac function (A, B, C). Therapeutic interventions can change the variables. Diuresis decreases preload (1), inotropes increase cardiac function at any given fill- ing pressure (2), and the use of beta or calcium channel blockers can inhibit contractility and move the patient’s cardiac function between curves I and II. standardized to body surface area (i.e., cardiac index = cardiac output/body sur- face area) facilitates the comparison of hemodynamic variables among patients. However, if the data is used primarily to predict trends over time, indexing pro- vides little added benefit. The principle on which the use of the PCWP as a measurement of left ventric- ular preload rests on is the assumption that inflation of the balloon in the wedge position within the pulmonary artery obstructs blood flow around the catheter tip and creates a static column of blood that is contiguous to the left atrium and, at end diastole before mitral valve closure, with the left ventricular end-diastolic pressure. Since catheter placement is flow-directed, the balloon usually carries the PAC tip to zone 2 of West, where the hydrostatic pressure in the pulmonary artery (Ppa) exceeds alveolar (PA) pressure, which exceeds pulmonary venous (Ppv) pressure (Ppa > PA > Ppv), or to zone 3 of West, where Ppa > Ppv > PA (Figure 2–6). Since the mean airway pressure in zones 1 and 2 is intermittently greater than pulmonary venous pressure, collapse of the vasculature causes inability to transduce accurate intravascular pressure. The reliability of the PCWP is greatest

40 The Intensive Care Manual FIGURE 2–6 West zones of the lung. West zones relate ventilation and perfusion. Flow is greatest in dependent zones, partially governed by gravity. Ventilation/perfusion ratio is greatest toward the apex of the lung. In order to best reflect cardiac function, the tip of the PAC should lie in zone 3 or 4. when the catheter tip is in zone 3 or 4, since only in these zones is there a contin- ual column of uninterrupted blood between the catheter tip and the left atrium. High mean airway pressure and hypovolemia are the most common causes of relatively decreased zones 3 and 4 of the lung. The risks and difficulties inherent in repeated efforts at repositioning guided by lateral chest radiographs is not practical, cost-effective, or safe. Instead, the catheter is assumed to be in zone 3 or 4, unless there is a marked transduction of pulmonary pressures on the pulmonary artery and PCWP waveforms. Airway PEEP increases the proportion of zones 1 and 2 relative to zone 3 be- cause of alveolar recruitment. The probability that the catheter tip lies in zone 3 or 4 is higher if a change in PCWP is less than half the incremental change in PEEP, and if the pulmonary artery diastolic (PAD) pressure is slightly higher than the PCWP. The true PCWP may be approximated using the formula: PCWP = PCWPM − 0.5(PEEP − 10)

2 / Intravascular Access and Hemodynamic Monitoring 41 Where PCWPM is the measured PCWP at any level of PEEP Because the PCWP, analogous to the CVP, reflects a balance between blood re- turn to the left side of the heart and the ejection of blood in the left ventricle by car- diac pumping function, it is not in itself an absolute. Elevated PCWP may indicate either fluid overload or decreased cardiac contractility. The PCWP does not reflect the volume of extracellular fluid. During myocardial ischemia, decreased ventricu- lar compliance and impaired contractility reduce the ability of the left side of the heart to maintain an effective forward flow of blood, which is reflected in decreased stroke volume and ejection fraction, causing the measured PCWP to become ele- vated at any given preload. The appropriate therapeutic intervention at this time, active decrease of preload or active increase in contractility, requires both more data regarding cardiac output and previous training and experience. The accuracy of the PCWP as an indicator of left ventricular end-diastolic pressure (LVEDP) is compromised in a number of pathophysiologic conditions in addition to the car- diopulmonary interactions described earlier.11 Mitral stenosis results in left atrial end-diastolic pressure and PCWP that are higher than LVEDP, an artifact caused by impaired left atrial ejection. The presence of “cannon” v waves on the pressure tracing can aid diagnosis of this condition. Large atrial masses, such as myxomas or mural thrombi, may falsely increase atrial pressures by decreasing atrial compliance and falsely elevate the PCWP. In aortic regurgitation, the PCWP underestimates the LVEDP because the mitral valve closes before left ventricular filling is completed. Regurgitant flow across the aortic valve continues to increase LVEDP and cannot be measured unless the mitral valve has also become incompetent. The CVP is always lower than the PCWP, except when pulmonary vascular resistance is substantially elevated, in which case CVP is higher than PCWP, or in the case of tamponade, in which the two pressures are equal. Pericardial tamponade restricts the filling of all cardiac chambers and results in the pathognomonic condition known as “equalization of pressures,” in which CVP, mean pulmonary artery pressure, and PCWP are equal. The PCWP and the pulmonary artery diastolic pressure are usually similar if the heart rate is less than 90 beats per minute. PULMONARY CAPILLARY WEDGE PRESSURE AND ADULT RESPIRATORY DISTRESS SYNDROME The PCWP is a useful guide to both pulmonary capillary filtration pressure and left ventricular filling pressure. The determination of the PCWP is often emphasized as a diagnostic tool in the differentiation of cardio- genic and noncardiogenic pulmonary edema. The diagnosis of ARDS rests on the PCWP determination; however, patients with ARDS are usually receiving venti- lation using PEEP. The relationship of actual and measured PEEP has been dis- cussed in the preceding section. Pulmonary capillary transmembrane fluid flux is described by Starling’s law, which defines the equilibrium between hydrostatic and osmotic forces across a capillary membrane:

42 The Intensive Care Manual F = [(Pi − Po) − (COPi − COPo )]× K Where F is transmembrane fluid flux Pi is hydrostatic pressure within artery PO is hydrostatic pressure outside the capillary COPi is intravascular oncotic pressure COPo is extravascular oncotic pressure K is the filtration coefficient or Pcap = PCWP + 0.4 (Pa − PCWP) Where Pcap is pulmonary capillary filtration pressure Pa is pulmonary artery pressure Elevation of the pulmonary capillary hydrostatic pressure in the presence of left ventricular failure favors transudation of fluid across the basement mem- brane and into the alveoli. When the volume of fluid overcomes the maximal lymphatic clearance, pulmonary edema occurs, manifested by a widening of the alveolar-arterial oxygen gradient and decreased lung compliance. Since a great many other factors can produce an identical picture (e.g., inflammation, high levels of negative alveolar pressure, hypoalbuminemia), the PCWP aids the dif- ferential diagnosis. Low or normal levels of PCWP in the presence of clinically determined pulmonary edema is a major criterion for the diagnosis of ARDS. CARDIAC OUTPUT MEASURED BY THERMODILUTION In addition to in- tracardiac pressure measurements, the PAC also enables the measurement of car- diac output by thermodilution.12 A thermistor at the tip of the PAC continually measures the temperature of the blood in the pulmonary artery as it flows past the catheter tip. Injection of a known quantity of fluid at a known temperature into the right atrium allows the change in the temperature of the mixed blood as it flows past the thermistor to be plotted as a function of time, as shown in Figure 2–7. The differentiated rate of change (dT/dt) is proportional and the integrated area under the curve is inversely proportional to the cardiac output. Thermodilu- tion is a modification of an indirect indicator dye (indocyanine green) dilution technique in which the flow is equal to the amount of dye injected divided by the integral of the instantaneous concentration of dye in sampled arterial blood over time. The determination of cardiac output using the Fick equation predates the PAC but nonetheless requires right heart catheterization. The Fick equation is:

2 / Intravascular Access and Hemodynamic Monitoring 43 FIGURE 2–7 Thermodilution cardiac output curves. The curves represent a change in tem- perature detected by the thermistor at the PAC tip as a mixed injectate of known temperature flows past. The curve with the greatest change in temperature (dT) per unit of time has the lower area under the curve but has the greatest cardiac output associated with it (A and B). Curve C depicts a thermodilution curve as sensed by a rapid response thermistor capable of determining end–diastolic volume (EDV) and subsequent volume changes (ESV) as the right ventricle empties. The ejection fraction (EF) is the EDV-ESV, which is the stroke volume SV, divided by EDV. CO = V˙ O2 CaO2 − CVO2 Where CO is cardiac output VO2 is whole body O2 consumption CaO2 is content of O2 in arterial blood CvO2 is content of O2 in venous blood This equation is now most commonly used as a method of calculating V˙ O2 when the cardiac output is measured directly, using the PAC. Thermodilution cardiac output determination is the currently accepted standard of practice and adds greatly to the utility of the PAC. Although room temperature injectates are reliable in most patients, the signal-to-noise ratio is more favorable with cold injectates, es- pecially in patients with low body temperature or low cardiac output. Variation of the cardiac output with phases of the respiratory cycle suggest that either measure- ments be timed to coincide with the same respiratory phase or an average value be used to predict trends. In clinical situations, trends which occur over time during the care of a patient are always more valuable than absolute numerical values.

44 The Intensive Care Manual The thermodilution cardiac output as determined by the PAC is subject to ar- tifactual inaccuracies based on the method used. Since the determination of car- diac output is directly based on the temperature change sensed by the thermistor at the catheter tip, smaller changes in temperature produce a falsely elevated level of cardiac output. Smaller temperature changes, which artifactually elevate the derived cardiac output, can occur with the use of less injectate than necessary or warmer-than-measured injectate (after a long wait at room temperature before injection of cold saline) and in the presence of right-to-left cardiac shunts. On the other hand, the most common cause of artifactually decreased cardiac output is tricuspid regurgitation, which allows a prolonged mixing time of blood and in- jectate, resulting in prolonged transit time in the right side of the heart and a de- crease of the temperature change with time. A rate of injection that is too slow also gives a falsely low value for cardiac output. Left-to-right shunts may make thermodilution cardiac output unmeasurable. Finally, rapid infusion of fluids may dilute the injectate and also render the cardiac output measurement inaccu- rate. DERIVED CARDIAC INDEXES Cardiac performance interpretation must be based on the fundamentals of cardiac physiology. The amount of blood ejected from each ventricle per heartbeat is the stroke volume (SV). The output of the left ventricle per unit time is the cardiac output (CO). CO can also be expressed as a function of heart rate and stroke volume: CO (L/min) = HR (beats/min) × SV (mL/beat) Where CO is cardiac ouput HR is heart rate SV is stroke volume Cardiac performance can then be expressed as a measure of the chronotropic and inotropic states of the heart. Chronotropy, or heart rate, is the effective bal- ance of vagal and sympathetic tone in the resting heart rate. Inotropy, or contrac- tility, is the sum of the tension generated by preload and the sum of contractility influences, including sympathetic effects on membrane receptors, channels, and intracellularly mediated contractility. However, CO also depends on the imped- ance, or resistance, to flow. This resistance is referred to as afterload and is low in the pulmonary arterial circulation right ventricular afterload and high in the aor- tic and systemic arterial circulation left ventricular afterload. Thus, afterload is most commonly referred to as either pulmonary vascular resistance (PVR) or systemic vascular resistance (SVR), although afterload is truly more complicated than vascular resistance alone. Although afterload is crucially important to car- diac performance, it can only be measured experimentally in heart-lung prepara- tions ex-vivo. The PAC cannot measure afterload or SVR but does allow an inference based on measured ventricular performance.

2 / Intravascular Access and Hemodynamic Monitoring 45 Mathematically, vascular resistance on the pulmonary and systemic circula- tions can be expressed as derivatives of Ohm’s law, which states that current is electromotive force divided by resistance, or flow equals pressure divided by re- sistance. Rearrangement to solve for vascular resistance produces: PVR = MAP − PCWP CO × 80 SVR = MAP − CVP CO × 80 Where MAP is mean arterial pressure CVP is central venous pressure CO is cardiac output Alternatively, ventricular afterload can be expressed as the myocardial wall tension during ejection as defined by the Laplace equation. Note that the CO and vascular resistances are thus mathematically inversely proportional. CO is mea- sured and vascular resistance is calculated, lending greater credence to the treat- ment of CO. A more direct estimate of aortic resistance is based on the relationship: R (aortic) = arterial pulse pressure SV Where R (aortic) is aortic resistance SV is stroke volume Note that the Poiseuille-Hagen formula suggests that resistance is also indi- rectly influenced by viscosity (hematocrit). Furthermore, patients with arteriove- nous shunting typically have decreased baseline SVR. Further manipulation of measured data can potentially increase the inferences possible PAC monitoring; however, these manipulations must be interpreted with caution. The data which is directly obtained from the PAC (i.e., CVP, PCWP, CO, SvO2)can be combined with ECG information and manipulated mathematically to derive additional in- dexes of hemodynamic function (Table 2–4). Note however, that since informa- tion that is directly measured, such as the CO, has greater validity than derived indices, such as SVR, the former should carry more weight in management deci- sions. ASSESSMENT OF CARDIAC PHARMACOLOGIC INTERVENTION The PAC is the gold standard for the clinical assessment of the physiologic response of the

46 TABLE 2–4 Hemodynamic Formulas and Normal Ranges Variable Formula Cardiac index (CI) CI (L /min/m2 ) = CO (L /min) BSA (m2 ) Stroke volume (SV) SV (mL per beat) = CO (L /min) HR (be Stroke index (SI) SI (mL per beat per m2 ) = SV (m B Systemic vascular resistance (SVR) SVR (dynes/sec/m2 − 5) =  MA  Pulmonary vascular PVR (dynes/sec/m2 − 5) =  MP resistance (PVR)  Left ventricular stroke LVSWI (g-m per beat per m2) = 0. work index (LVSWI) (mL per beat per m2) NOTES: Units are in parentheses in equations. ABBREVIATIONS: BSA, body surface area; CO, cardiac output; HR, mean pulmonary artery pressure; PCNP, pulmonary capillary w

) × 1000 (mL /L) Normal Range eats per min) 2.8–4.2 L/min/m2 60–90 mL per beat mL per beat) 30–65 mL per beat per m2 BSA (m2 ) 1200–1500 dynes/sec/m2 100–300 dynes/sec/m2 AP (mm Hg) − CVP (mm Hg) × 80  45–60 CO (L /min) PAP (mm Hg) − PCWP (mm Hg) × 80  CO (L /min) .0136 [MAP (mm Hg) − PCWP (mm Hg)] SI , heart rate; MAP, mean arterial pressure; CVP, central venous pressure; MPAP, wedge pressure.

2 / Intravascular Access and Hemodynamic Monitoring 47 critically ill patient to therapeutic intervention (Figure 2–5). Because of this mode of assessment, pharmacologic intervention that affects cardiac perfor- mance can be specifically, and often selectively, directed at preload, chronotropy, inotropy, or afterload. Preload is increased by the administration of fluids that replenish or expand intravascular volume, such as blood products, colloids, or crystalloid (Figure 2–5a,b,c). Effective decreases in preload can be accomplished relatively by veno- dilating agents, such as low-dose nitrates or morphine, or definitively through diuresis (Figure 2–5, arrow 1). Chronotropy can be increased by vagolytic agents, such as atropine sulfate or related compounds, indirect and direct sym- pathomimetic agents, or artificial electrical pacing. Indirect sympathomimetic agents are those compounds, such as ephedrine, which trigger the release of epi- nephrine from sympathetic nerve terminals, and direct sympathomimetic agents are the epinephrine analogs, such as isoproterenol, which acts directly on the β1-receptors to increase heart rate. Heart rate is also indirectly regulated by the carotid baroreceptors and possibly also by atrial stretch receptors (Bainbridge re- flex). Inotropy can be decreased (Figure 2–5, arrow 3) indirectly through the blockade of β1-receptors or calcium antagonists and directly through depression of excitation-contraction coupling at the subcellular level. Recently, the identifi- cation and demonstration of physiologically active myocardial β3-receptors13 that exert negative effects on the inotropic state of the human heart have opened a new and exciting potential avenue of therapeutics based on the selective stimula- tion and blockade of this receptor. Inotropy can likewise be augmented (Figure 2–5, arrow 2) with indirect and direct β1-receptor stimulation by means of sympathomimetic agents, with inhibition of the membrane-based transtubular sodium-potassium ATPase pump by means of digitalis glycosides (which in- crease calcium flux into the myocytes), with manipulation of the serum-ionized calcium concentration relative to intracellular calcium concentration by means of administration of intravenous calcium salts, and by means of the inhibitors of phosphodiesterase (PDE), such as aminophylline, and specifically the inhibition of PDE-3 by amrinone and milrinone. Afterload can be increased by the general- ized stimulation of sympathetic tone, administration of vasopressin, or by selec- tive activation of α1-receptors, which precipitate vasoconstriction. Cold-induced vasoconstriction and increased cardiac afterload are an often-unrecognized cause of increased cardiac workload and therefore a potential cause of cardiac ischemia. However, afterload can be decreased pharmacologically by activators of the nitric oxide pathway, such as sodium nitroprusside; calcium channel blocking agents, such as nicardipine; direct smooth-muscle dilators, such as hydralazine; in- hibitors of angiotensin-converting enzyme (ACE), such as captopril; or indirect sympathectomy, affecting central sympathetic outflow. ASSESSMENT OF CARDIOPULMONARY INTERACTION Cardiac and pul- monary function are highly interdependent, and changes in pleural and in- trathoracic pressure, oxygenation, and ventilation exert important effects on pulmonary blood flow and left-sided CO.14 During spontaneous ventilation, the

48 The Intensive Care Manual generation of negative pleural pressure aids in thoracic venous return and aug- ments cardiac diastolic filling. The implementation of positive pressure ventila- tion by definition prevents the generation of negative intrathoracic pressure. Incremental increases in intrathoracic pressure (e.g., PEEP, mean airway pres- sure, peak airway pressure) progressively impair the venous return to the right side of the heart (preload) and thereby decrease CO. Transmitted pleural pres- sure to the compliant right side of the heart diminishes distensibility during dias- tole and further impairs venous return. Simultaneously, alveolar distention can impinge on pulmonary capillary flow, increase PVR and impose increased after- load. Bowing of the ventricular septum into the left ventricle occurs late as pres- sures in the right side of the heart increase further. However, most patients who require high levels of positive airway pressure to maintain adequate oxygenation have alveolar hypoxia, which may increase PVR through hypoxic pulmonary vasoconstriction, further impairing function of the right side of the heart. Conversely, increasing alveolar recruitment and resolution of alveolar hypoxia may, with continued application, improve cardiac function by diminishing the presence of hypoxic pulmonary vasoconstriction. PEEP may decrease left ventricular afterload and, until decreased preload intervenes, tran- siently increase CO. Since pressures in the right side of the heart with high levels of positive pressure ventilation may not be a reliable indicator of preload, the measurement of CO, mixed venous saturation, or right ventricular ejection frac- tion is almost always considered a mandatory intervention. CONTINUOUS CARDIAC OUTPUT AND MIXED VENOUS OXIMETRY CO can be measured at 2- to 3-minute intervals by a catheter that includes a thermal filament at the level of the right ventricle. CO determinations at frequent regular intervals allow “continuous” CO monitoring. In addition, injectate fluid load and operator variability are reduced and changes in physiologic state more rapidly detectable. CO is proportional to the saturation of mixed venous blood; the higher the CO the greater the saturation of mixed venous blood. Fiberoptic technology al- lows some PACs to measure the saturation of hemoglobin in the blood as it flows past the PAC tip in the pulmonary arteries, using the principle of reflectance spectrophotometry. Mixed venous blood is blood that includes blood return from all organs, including the heart via the thebesian veins, just before reoxy- genation in the pulmonary capillaries. In catheters that are not equipped with fiberoptic systems capable of measuring the hemoglobin saturation directly, blood can be slowly aspirated from the distal port of the PAC and submitted for ABG analysis. Sepsis causes a decrease in both peripheral V˙ O2 and arteriovenous shunting, which is reflected as an increase in the SvO2. Variables that contribute to DO2 (e.g., anemia, hypovolemia, cardiogenic shock, arterial hypoxemia, car- boxyhemoglobinemia) all cause a decrease in the SvO2. Pathologic left-to-right intracardiac shunting produces an increase in the amount of oxygenated blood in the right ventricle and thus an increase in the SvO2 and a grossly inaccurate (arte- riovenous) O2 gradient.

2 / Intravascular Access and Hemodynamic Monitoring 49 PULMONARY ARTERY CATHETERS AND VOLUMETRIC RIGHT VENTRIC- ULAR EJECTION FRACTION A variant of the conventional PAC uses a rapid re- sponse thermistor, capable of determining right ventricular end-systolic volumes (ESV) and end-diastolic volumes (EDV), and thereby makes possible the calcula- tion of the right ventricular ejection fraction (RVEF) (Figure 2–7c), expressed as RVEF = ESV/EDV.15 This particular catheter therefore combines measured volu- metric and pressure data to increase the sensitivity and reliability of the PAC in the assessment of biventricular function. There is evidence to suggest that patients admitted to the ICU with significant right ventricular dysfunction that cannot be predicted with conventional PAC monitoring.16 The diagnostic superiority of right ventricular end-diastolic vol- ume (RVEDV) measurement over measurements of urine output, CVP, and PCWP has been repeatedly suggested in patients with burns, sepsis, and trauma. The effect of high positive airway pressures, especially PEEP, disparately affects the thin walled right ventricle. The effect of increasing PEEP on right ventricular afterload and the coincident increase in RVEDV and depression of RVEF is best assessed by using the volumetric catheter. Filling pressures thus are more likely to be overestimated in the presence of undetected right heart dysfunction. Advan- tages to RVEF-based cardiac performance evaluation has not been demonstrated in cardiac patients. The RVEF catheter loses accuracy in the presence of arrhyth- mias or tricuspid valve regurgitation. OXYGEN KINETICS AND PULMONARY ARTERY CATHETERS One of the most important applications of the PAC is information regarding whole body oxygen delivery (D˙ O2) and volume of oxygen uptake (V˙ O2). The caveat that D˙ O2 and V˙ O2 are systemic indexes not applicable to individual tissue beds should not detract from their utility as gross measures of the adequacy of resuscitation. The primary determinant of D˙ O2 is cardiac output; therefore, an incremental change in cardiac output (CO) is more important to D˙ O2 than an equal incremental change in CaO2 or its comprised variables. D˙ O2 = CaO2 × CO D˙ O2 ={((Hb)× 1.34 × SaO2) + (PaO2 × 0.0031 )} × CO Oxidative phosphorylation is a more efficient pathway for substrate metabo- lism than anaerobic glycolysis and generates more adenosine triphosphate (ATP) per mole of glucose, 36 moles in comparison to 2 moles, respectively. ATP pro- vides cellular bioenergy for enzymatic pathways. For oxidative phosphorylation to predominate, D˙ O2 must at least match V˙ O2. Oxygen delivery that is inadequate to meet metabolic demand, or the metabolic requirement for oxygen (MRO2), defines the state of shock. Since D˙ O2 depends primarily on cardiac output (CO), shock states are classified based on the underlying pathophysiologic mechanism that causes the compromise in CO. Progressive mismatches between D˙ O2 and V˙ O2 are manifested first by a decreasing mixed venous oxygen saturation, an excess of systemic lactate, and metabolic acidosis. Under normal conditions D˙ O2 exceeds

50 The Intensive Care Manual V˙ O2 significantly, and there is a range over which decreases in D˙ O2 have no de- tectable metabolic consequences. Compensation in this range occurs through modulation of the oxygen extraction ratio (O2ER). The O2ER is the ratio of V˙ O2 to D˙ O2 and represents the fraction of delivered oxygen that is taken up into the tissues, usually in the range of 20% to 30%. O2ER = V˙ O2 D˙ O2 ×100 Maximal oxygen extraction occurs when the O2ER approximates 50% to 60%. At this point, known as the point of critical oxygen delivery, the VO2 becomes de- pendent on DO2 (supply-dependent VO2), a state also known as dysoxia. Lactate production increases progressively as this mismatch increases. This can be de- picted in a relationship (Figure 2–8) that has great conceptual utility,17 but re- mains controversial despite many years of application.18 However, since plasma lactate level represents a balance between production and extraction in tissues, such as the liver, lactate may not be detectable despite increases in production. Lactate excess may also be a result of endotoxin inhibition of pyruvate dehydro- FIGURE 2–8 The theoretical relationship between oxygen delivery (D˙ O2) and oxygen uptake (V˙ O2): Gradual decrease in D˙ O2 has little or no detectable effect on V˙ O2 since compensation occurs by icnreased peripheral extraction. Further decrease in D˙ O2 to the inflection point causes the V˙ O2 to become pathologically flow dependent and a state of dysoxia occurs in which there is a change from oxidative to anaerobic metabolism. Pathologic supply depen- dency is heralded by the development of lactic acidosis. Although theoretically useful, the rela- tionship is controversial because mathematical coupling can occur when both V˙ O2 and D˙ O2 are measured using the same device (PAC), and it does not account for possible metabolic al- terations induced by inflammatory mediators which may alter V˙ O2 at the tissue and cellular levels.

2 / Intravascular Access and Hemodynamic Monitoring 51 genase and thiamine deficiency.19 The point of critical oxygen delivery is proba- bly much higher than normal in critically ill patients, possibly because the tissue MRO2 level is increased by metabolic stress. ECHOCARDIOGRAPHIC AS]SESSMENT OF CARDIAC FUNCTION More recently, minimally invasive and noninvasive measures of cardiac perfor- mance have achieved some measure of popularity. Transesophageal echocardiog- raphy (TEE) is the best established of these and has greater applicability in the ICU patient because of thoracic pathophysiology, which often limits the size of acoustic windows available to transthoracic echocardiography (TTE).20 Esophageal or gastric placement of the ultrasound TEE transducer results in close proximity of the transducer to the heart and, therefore, minimal image degrada- tion by air interfaces. TEE in two and three dimensions (2-D, 3-D) provides real- time visualization of ventricular dimensions in diastole and systole, Doppler imaging of flow, and computer-assisted calculation of ejection fraction.21 Doppler imaging has great value in the imaging of valvular heart disease, pul- monary blood flow, hepatic blood flow, and intracardiac shunts and can be either color-flow Doppler or pulsed Doppler. TEE is extremely valuable in the diagnosis of mechanical obstruction of cardiac function, such as pericardial fluid and tam- ponade, atrial myxoma, pulmonary embolus, and prosthetic valve failure. TEE may also aid in the early bedside evaluation of suspected thoracic artery aneurysm or dissection, as an adjunct to arteriography. In addition, segmental wall-motion abnormalities are significantly more sensitive and earlier indicators of ischemia than are changes in the PCWP or the ECG. The value of adjunctive information obtained echocardiographically in the specific setting of the ICU is well documented and TEE skills are important, if not vital, tools for the hemody- namic management of critically ill patients. THORACIC BIOIMPEDANCE PLETHYSMOGRAPHY AND ESOPHAGEAL DOPPLER TECHNOLOGY The use of thoracic impedance plethysmography is a noninvasive alternative to using invasive vascular cannulation solely for the purposes of diagnostic mea- surements. This technique is based on changes in the electrical impedance of the thoracic cavity that occur with changes in thoracic blood volume throughout the cardiac cycle. An alternating current of small amplitude (2.5 to 4.0 mA) traverses the chest at a frequency of 70 to 100 kHz. Four (transmitter-sensor) pairs of cuta- neous electrodes determine the impedance to current flow. Since respiratory variations occur at a lower frequency than cardiac variations, the effect of respi- rations can be eliminated. Stroke volume (SV) is determined mathematically, based on the specific resistivity of blood, thoracic length, basal thoracic imped- ance, ventricular ejection time, and the maximum rate of impedance change dur-

52 The Intensive Care Manual ing systolic upstroke. The SV correlates with the impedance change over a car- diac cycle. Cardiac output can be readily derived from the SV by the equation CO = SV × HR, where HR is heart rate. Additional indices, such as ejection velocity index, ventricular ejection time, and thoracic fluid index, can assist in more sub- tle evaluation of cardiac function. Since bioimpedance estimates the pulsatile component of SV, conditions in which the flow is more continuous than pul- satile (e.g., sepsis, hemodilution) may artifactually lower the calculated CO. The use of bioimpedance plethysmography is very limited in the presence of arrhyth- mias. Since plethysmography requires an estimated CVP, its utility can be greatly increased if the CVP is directly measured by an invasive method. Thoracic bioimpedance overestimates CO: (1) if the CVP is lower than estimated, (2) in the presence of low cardiac flow, (3) when inotropes are used, and (4) in the presence of aortic insufficiency. Thoracic bioimpedance underestimates CO in the presence of sepsis, hypertension, and intracardiac shunts. In general, bioim- pedance has utility in the intermittent, short-term, or initial evaluation of cardio- vascular function in the critically-ill patient but is severely limited in continuous and intensive long-term management. An alternative to thoracic bioimpedance plethysmography is transesophageal Doppler monitoring. Monitoring of cardiac function with Doppler technology estimates the velocity of blood flow in the descending aorta and mathematically derives correlates of CO and afterload. The esophageal doppler may incorporate M-mode ultrasound technology to standardize the placement of the doppler probe with respect to aortic diameter. Since the esophageal Doppler method re- lies mainly on direct measurement and less on assumption, the thoracic Doppler probe shows promise in the care of critical care patients. MUCOSAL TONOMETRY The mucosal tonometer is a potentially useful method for the assessment of tissue-specific perfusion. Tonometer technology has been applied to the esoph- ageal and gastric mucosa, the mucosa of the rectum, and the sublingual oral mu- cosa. Although the monitoring site varies, the technology and the physiologic principles are the same. The central role of the gastrointestinal tract postulated in the gut motor hypothesis is as an initiator and perpetuator of bacteremia and inflammatory mediator release.22 The splanchnic circulatory system is among the first to be affected by systemic shock, and splanchnic hypoperfusion continues for a time after systemic variables (e.g., blood pressure, HR, urine output, lactic acid level, CO) have been restored to normal.23 Local acidosis resulting from hy- poperfusion and dysoxia can be quantitatively estimated using the Henderson- Hasselbach equation, the measured arterial bicarbonate, and the CO2 inside a silastic balloon, which theoretically reflects tissue PCO2. Although mucosal tonometry represents a technology in evolution and is therefore controversial, it has tremendous potential as a measurable and tissue-specific endpoint for resus- citation.

2 / Intravascular Access and Hemodynamic Monitoring 53 pHi = 6.1 + log10 arterial HCO3 saline PCO2 × 0.03 Where pHi = intracellular pH HCO3 = bicarbonate saline PCO2 = partial pressure of carbon dioxide in the saline balloon CAPNOGRAPHY Although capnography is readily available in the CCU, its application as an ad- junctive tool for hemodynamic monitoring is not well appreciated. The capno- graph detects, measures, and depicts the respiratory flow of CO2 during expiration. The presence of CO2 in exhaled gas is an indicator of ventilation (and endotracheal tube placement) and pulmonary perfusion. Quantitative measure of the concentration of CO2 at end-expiration, known as the end-tidal CO2 (CO2ET), is both a measure of ventilation and perfusion. In instances where per- fusion is thought to be constant, titration of ventilation to the CO2ET can opti- mize ventilation. However, during steady-state ventilation, alterations in the CO2ET signify changes in cardiac output and resultant changes in pulmonary perfusion. In fact, the presence of CO2ET in the early phases of cardiac resuscita- tion are associated with improved outcome. SUMMARY A fundamental indication for admission to the ICU is the availability of special- ized technology and personnel to facilitate rapid intervention for diagnosis and therapy. The cost of monitoring must be continuously weighed against the potential benefit and risk. Costs may be direct, such as acquisition costs of equipment, or indirect, such as the additional personnel time that must be committed to main- tenance, data acquisition and recording, and treatment of complications. If ex- pensive technology cannot be demonstrated to positively affect patient outcome, justification of its use becomes increasingly difficult. To a large extent, it is vital to recognize that monitoring provides data that only becomes meaningful when properly interpreted and used for timely and appropriate intervention. REFERENCES 1. Gosbell IB. Central venous catheter-related sepsis: Epidemiology, pathogenesis, diag- nosis, treatment and prevention. Int Care World 1994;11:54.

54 The Intensive Care Manual 2. Kamal GD, Pfaller MA, Rempe LE, et al. Reduced intravascular catheter infection by antibiotic bonding: A prospective, randomized, controlled trial. JAMA 1991;265:2364. 3. Curtas S, Tramposch K. Culture methods to evaluate central venous catheter sepsis. Nutr Clin Pract 1991;6:43. 4. Randolph AG, Cook DJ, Gonzales CA, et al. Benefit of heparin in central venous and pulmonary artery catheters: A meta-analysis of randomized clinical trials. Chest 1998;113:165. 5. Seneff M. Arterial line placement and care. In Rippe JM, Irwin RS, Fink MP, eds. Pro- cedures and techniques in intensive care medicine. Boston: Little, Brown, 1995:15. 6. Kleinman B, Powell S, Kumar P, et al. The fast flush test measures the dynamic re- sponse of the entire blood pressure monitoring system. Anesthesiology 1992;77:1215. 7. Iberti TJ, Fischer EP, Leibowitz AB, et al. A multicenter study of physicians’ knowl- edge of the pulmonary artery catheter. JAMA 1990;264:2928. 8. Connors AF, Speroff T, Dawson NV, et al. The effectiveness of right heart catheteriza- tion in the initial care of critically ill patients. JAMA 1996;276:889. 9. American Society of Anesthesiologists task force on pulmonary artery catheterization. Anesthesiology 1993;78:380. 10. Pulmonary Artery Catheter Consensus Conference: Consensus statement. Crit Care Med 1997;25:10. 11. Tuman KJ, Carroll GC, Ivankovich AD. Pitfalls in interpretation of pulmonary artery catheter data. J Cardiothorac Anesth 1989;3:625. 12. Nishikawa T, Dohi S. Errors in the measurement of cardiac output by hemodilution. Can J Anaesth 1993;40:142. 13. Bond RA, Lefkowitz RJ. The third beta is not the charm. J Clin Invest 1996;98:241. 14. Diebel LN, Myers T, Dulchavsky S. Effects of increasing airway pressure and PEEP on the assessment of cardiac preload. J Trauma 1997;42:585. 15. Ivatury RR, Simon RJ, Islam S, et al. A prospective randomized study of end points of resuscitation after major trauma: Global oxygen transport indices versus organ- specific gastric mucosal pH. J Am Coll Surg 1996;183:145. 16. Hoffman MJ, Greenfield LJ, Sugerman HJ, et al. Unsuspected right ventricular dys- function in shock and sepsis. Ann Surg 1983;198:307. 17. Shoemaker WC, Appel PL, Krom HB. Role of oxygen debt in the development of organ failure, sepsis, and death in high-risk surgical patients. Chest 1992;102:208. 18. Ronco JJ, Phang PT, Walley KR, et al. Oxygen consumption is independent of changes in oxygen delivery in severe adult respiratory distress syndrome. Am Rev Respir Dis 1991;143:1267. 19. Mizock BA, Falk JL. Lactic acidosis in critical illness. Crit Care Med 1992;20:80. 20. Porembka DT, Hoit BD. Transesophageal echocardiography in the intensive care pa- tient. Crit Care Med 1991;19:826. 21. Feinberg MS, Hopkins WE, Davila-Roman VG, et al. Multiplane transesophageal echocardiographic doppler imaging accurately determines cardiac output measure- ments in critically ill patients. Chest 1995;107:769. 22. Aranow JS, Fink MP. Determinants of intestinal barrier failure in critical illness. Br J Anaesth 1996;77:71. 23. Fiddian-Green RG. Gastric intramucosal pH, tissue oxygenation, and acid base bal- ance. Br J Anaesth 1995;74:591.

CHAPTER 3 Approach to Shock PETER J. PAPADAKOS “In acute diseases, coldness of the extremities is a very bad sign.” HIPPOCRATES, 400 BC INTRODUCTION DIFFERENTIAL DIAGNOSIS PATHOPHYSIOLOGY Distributive Shock Hypovolemic Shock GENERAL PRINCIPLES Obstructive Shock Cardiogenic Shock EVALUATION OF SYMPTOMS History MANAGEMENT AND THERAPY Hypovolemic Shock Distributive Shock Fluid Management Obstructive Shock Vasoactive Drugs Monoclonal Antibodies DIAGNOSTIC TESTING Coagulation SUMMARY Hematologic Parameters Renal Parameters Echocardiography COMPLICATIONS OF SHOCK 55 Copyright 2001 The McGraw-Hill Companies. Click Here for Terms of Use.

56 The Intensive Care Manual INTRODUCTION Shock in its various forms presents a very common challenge in the ICU. Over the past decade, we have developed new terminology to better understand shock. In an attempt to develop a common language, a consensus conference of the So- ciety of Critical Care Medicine and the American College of Chest Physicians was held in August 1991 to produce a series of universal definitions for the systemic inflammatory response syndrome (SIRS).This interplay of cellular and systemic responses, which may be modulated by cytokines, may replace the terms septic, cardiogenic, hypovolemic, distributive, and obstructive shock. This enhanced understanding of the pathophysiology of shock syndromes and SIRS may not only give us a common language but also may aid in the develop- ment of treatment protocols. We now understand that the immediate recogni- tion of and institution of treatment for SIRS and shock are paramount in the ICU. PATHOPHYSIOLOGY The term “shock” can simply be defined as inadequate tissue perfusion along with cellular hypoxia and oxygen debt, which results in cellular dysfunction and is caused by inadequate systemic oxygen delivery or impairment of cellular oxy- gen uptake. This can be the result of poor oxygen delivery, maldistribution of blood flow, the effect of cytokines on cell function, a low perfusion pressure, or a combination of these factors.1–3 The common denominator in all shock states and the earliest manifestation of shock is reduced oxygen consumption (VO2). Cellular hypoxia can incite SIRS and multiple organ dysfunction syndrome (MODS). It is obvious that oxygen debt should be rapidly reversed and systemic oxygen delivery and consumption maintained. The oxygen debt is caused by a low flow in hypovolemic or cardiogenic shock, by a cellular or metabolic deficit in septic shock, and by a maldistribution of blood flow in other types of shock. This decade has lead to an understanding of how various cytokines may be released during shock states and interplay with various organ systems to cause end-organ damage or MODS. Cytokines, for example, nitric oxide (NO), interleukin-2 (IL-2), and tumor necrosis factor (TNF), and their release in vari- ous forms of shock may modulate the microvascular and cellular responses of shock. The low flow, or hypovolemia, in some forms of shock may also be re- sponsible for the release of various cytokines. The cytokine cascade in SIRS may account for many of the presenting signs of shock, such as respiratory failure, capillary leak, shunting, redistribution, depressed myocardial function, oxygen uncoupling, and cellular ischemia. The mediator response in SIRS can be divided into four phases based on the cytokine-cellular response:

3 / Shock 57 1. Induction 2. Triggering of cytokine synthesis 3. Evolution of cytokine cascade 4. Elaboration of secondary mediators with ensuing cellular injury The events following endotoxin exposure provide a good model for discussion of the phases of SIRS. Endotoxin, shed from bacteria as they multiply or die, is one of the most powerful triggers of SIRS and acts by stimulating phagocytic cells, particularly macrophages, to synthesize TNF-α, which then activates the complement coagulation cascades and also induces endothelial cell activation. GENERAL PRINCIPLES The general progression of the SIRS response may be similar or different for each type of shock, and the use of hemodynamic profiles may aid in its classification (Table 3–1). Although certainly not mandated as a management tool, data gath- ered using a pulmonary artery catheter (PAC) facilitates classification and under- standing of the causes of shock. Furthermore, knowledge of these hemodynamic profiles is helpful in the diagnosis and management of shock. New monitors, such as esophageal echocardiographic and esophageal Doppler ultrasonographic hemodynamic monitors, HemoSonic 100 (Arrow International Reading, PA) may give a clinician real-time data on volume status and cardiac contractility, which may eventually prove more helpful than the PAC. More specific symptoms for each type of shock are shown in Table 3–2. Some older terminology, such as hyperdynamic shock (an increased cardiac output and lowered vascular resistance), early and late shock, and warm and cold shock, are no longer used. The body’s responses to shock may vary according to the cause. For example, distributive shock may be characterized by low cardiac output. This variability may progress in all forms of shock to a set of common organ effects (Table 3–3). TABLE 3–1 Hemodynamic Profiles of Various Types of Shock Type of Shock Pulmonary Artery Cardiac Output SVR Occlusion Pressure ↑ Cardiogenic ↑ ↓ ↑ Hypovolemic ↓ ↓ ↓ Distributive ↓ or nl ↑ or nl or ↓ ↑ Obstructive ↑ or nl or ↓ ↓ ABBREVIATIONS: SVR, systemic vascular resistance; ↑, increases; ↓, decreases; nl, normal.

58 The Intensive Care Manual TABLE 3–2 General Symptoms of Shock CNS Changes • Confusion • Coma • Combative behavior • Agitation • Stupor Skin Changes • Cool • Clammy • Warm • Diaphoresis Cardiovascular • Increase or decrease in heart rate • Arrhythmia • Angina • Low, high, or normal cardiac output • Changes in pulmonary pressure (see Table 3–1) Pulmonary • Increased respiratory rate • Increase or decrease in end-tidal CO2 • Decrease in O2 saturation • Increased pulmonary pressures • Respiratory failure • Decreased tidal volume • Decreased FRC Renal • Decreased urine output • Elevation in BUN and creatinine levels • Change in urine electrolyte levels ABBREVIATIONS: CNS, central nervous system; CO2, carbon dioxide; O2, oxygen; FRC, functional resid- ual capacity; BUN, blood urea nitrogen. EVALUATION OF SYMPTOMS History The type of shock must be evaluated by reviewing the history of the disease process. In cardiogenic shock, the patient may have a history of cardiac disease, poor cardiac function, congestive heart failure, myocardial ischemia, or valvular heart disease. In hypovolemic shock, there is usually a history of blood loss, trauma, fluid losses, dehydration, third spacing, or other fluid losses. Distributive shock is usually associated with exposure to an infectious or allergic agent, neurologic events, or a reaction to various immunologic substances. In obstructive shock,

3 / Shock 59 TABLE 3–3 Common Effects of Shock on Organs Systemic • Capillary leak • Formation of microvascular shunts • Cytokine release Cardiovascular • Circulatory failure • Depression of cardiovascular function • Arrhythmia Hematologic • Bone marrow suppression • Coagulopathy • Disseminated intravascular coagulation (DIC) • Platelet dysfunction Hepatic • Liver insufficiency • Elevation of liver enzyme levels • Coagulopathy Neuroendocrine • Change of mental status • Adrenal suppression • Insulin resistance • Thyroid dysfunction Renal • Renal insufficiency • Change in urine electrolyte levels • Elevation of BUN and creatinine levels Cellular • Cell-to-cell dehiscence • Cellular swelling • Mitochondrial dysfunction • Cellular leak there may be a history of trauma or a process that leads to a mechanical obstruc- tion of cardiac filling, such as cardiac tamponade. Early recognition of hypotension and hypoperfusion is essential in prevention and treatment of all types of shock. Hypoperfusion may be the trigger for much of the end-organ dysfunction and cytokine activation. In adults, a drop in sys- tolic blood pressure of more than 40 mm Hg constitutes significant hypotension. Hypoperfusion may be present in the absence of significant hypotension if mi- crocirculatory factors are activated. Shock is usually recognized as hypotension characterized by hypoperfusion abnormalities.

60 The Intensive Care Manual General symptoms are illustrated in Table 3–2 and are a guide for rapid evalu- ation and treatment. Hypovolemic Shock Hypovolemic shock occurs when there is a depletion of fluid in the intravascular space as a result of hemorrhage, vomiting, diarrhea, dehydration, capillary leak, or a combination of these. Capillary leak is common with the activation of the systemic inflammatory response.1 The hemodynamic findings in hypovolemic shock are decreased cardiac output, decreased pulmonary capillary wedge pressure (PCWP), and an increase in systemic vascular resistance (SVR). The echocardiographic and echodoppler profile is one of decreased right-sided filling, decreased stroke volume, and decreased aortic diameter. Distributive Shock The most common cause of distributive shock is septic shock. Other forms of distributive shock are anaphylactic shock, acute adrenal insufficiency, and neuro- genic shock. The primary problems are the development of shunts and capillary leak. In distributive shock, there is activation of SIRS and a breakdown of cellular function in the septic process. The hemodynamic profile is characterized by a normal or increased cardiac output with a low SVR and low-to-normal left ven- tricular filling pressure. The echocardiographic profile is one of low stroke vol- ume and an increase in aortic diameter. Obstructive Shock Direct mechanical obstruction to cardiac filling is the keystone of obstructive shock. In obstructive shock, there is depression of the ability to fill the right side of the heart, which may be the result of a fluid collection around the heart, car- diac tamponade, or a massive increase in intrathoracic pressure. In cardiac tam- ponade, the pressure in the right side of the heart, the pulmonary artery, and the left side of the heart equilibrate in diastole. A drop of more than 10 mm Hg in systolic blood pressure during inspiration, or paradoxical pulse, is an important finding.4,5 Another form of obstructive shock is tension pneumothorax, in which there is increased intrathoracic pressure with hypotension, resulting from de- creased preload. DIAGNOSTIC TESTING General laboratory tests should include measurement of blood lactate level (usu- ally secondary to anaerobic metabolism), which is a marker for poor oxygen de- livery or use6 and serum bicarbonate level (a decrease in this level is a marker for

3 / Shock 61 metabolic acidosis). There can also be an elevation in blood glucose level and changes in the level of several electrolytes: zinc, magnesium, and calcium; these should be measured.7 Alterations in renal parameters commonly include an ele- vation in creatinine and blood urea nitrogen (BUN) levels and changes in urine electrolyte levels. Liver parameters are also affected by shock states; alterations occur in all liver enzyme levels. ABG analysis is one of the most important labo- ratory tests because it measures the baseline oxygen delivery and utilization, which is the basic problem in shock. The most common findings are hypoxia, metabolic acidosis, and an elevation in PaCO2. Coagulation The coagulation cascade may be affected by the shock syndrome through activa- tion of SIRS with evidence of disseminated intravascular coagulation (DIC), an increase in fibrin split products, and a fall in fibrinogen and antithrombin III lev- els. Coagulation factors are also affected by liver failure, with increases in pro- thrombin time (PT) and activated partial thomboplastin time8 (APTT). Hematologic Parameters In septic or infectious shock, the WBC count can be either high or low. In other forms of shock, bone marrow suppression may lead to decreased production of all hematologic cells. Platelet counts may fall or platelets may not function nor- mally in several forms of shock. Erythropoietin levels also decrease in shock. Renal Parameters Oliguria and renal insufficiency are important markers for shock because the kidney is very sensitive to hypoperfusion and cytokine effects.9 Oliguria may be caused by direct renal injury by cytokines, prerenal volume problems, or post- renal problems. In all critically ill patients, a urine output of less than 0.5 mL/kg per hour is defined as oliguria. Echocardiography The addition of echocardiography in the ICU has added greatly to our ability to diagnose and manage various forms of shock. Formal echocardiography requires special training for both the transthoracic and esophageal forms. Over the past few years, an esophageal Doppler echocardiographic probe has been developed that is easy to use and gives data on aortic artery diameter, stroke volume, and cardiac output in real time.10

62 The Intensive Care Manual COMPLICATIONS OF SHOCK The most serious complication of shock is that low tissue perfusion may be an activating factor of SIRS through the release and modulation of cytokines and other vasoactive substances. This low-flow state and cytokine modulation may be at the heart of organ failure and MODS. The two most sensitive organs are the lungs (at risk for ARDS) and the kidneys (at risk for acute renal failure, or ARF). If this low-flow state is not rapidly corrected, other organ dysfunction occurs.1,2 Metabolic acidosis or lactic acidosis is a sensitive marker for low-flow states, and prolonged elevations of serum lactate level may be markers for increased mor- bidity and mortality.6 The cytokine cascade not only modulates vascular tone, but also controls other physiologic functions, such as bone marrow production, cell permiability, and electrolyte regulation, as well as DIC. DIFFERENTIAL DIAGNOSIS Distributive Shock Sepsis, accompanied by capillary leak, shunting, and microvascular changes, is the classic example of distributive shock. Sepsis is a form of distributive shock that occurs as a complication of a severe infection. The various circulatory and cellular events are caused by systemic activation of the inflammatory cascade and release of numerous mediators from tissue, mast cells, and circulating basophils. Another cause of distributive shock is anaphylaxis, an immediate hypersensitiv- ity reaction, which is mediated by the interaction of immunoglobulin (IgE) anti- bodies on the surface of most white cells and basophils. Anaphylaxis can be triggered by drugs, especially antibiotics (e.g., beta-lactamase inhibitors, cephalosporins, sul- fonamides), and animal toxins (from the stings of hornets, wasps, and bees). Other causes include heterologous serum, such as tetanus antitoxin, snake antitoxin, serums, blood transfusions, immunoglobulins, and vaccine products. Many health care workers and patients are now allergic to latex, which is found in countless prod- ucts used in health care, industrial preparations, and in the home. There have been multiple deaths and these events are now commonly reported by the lay press; spe- cial care must be taken in caring for these patients because latex is commonly used in gloves, IV tubing, and many other products. Anaphylactoid reactions are very similar to anaphylaxis, but without activa- tion of IgE, and can be caused by a wide range of materials and agents, including ionic contrast media, protamine, opioids, polysaccharide volume expanders (e.g., dextran, hydroxyethyl starch), anesthetics, and muscle relaxants. Neurogenic shock is another form of distributive shock. It involves loss of pe- ripheral vasomotor control as a result of neurologic dysfunction or injury to the nervous system. Adrenal gland dysfunction can also trigger distributive shock. Adrenal crisis can be caused by a deficiency of adrenal production of mineralocorticoids and

3 / Shock 63 glucocorticoids. It can be triggered or caused by adrenal hemorrhage, trauma, or overwhelming infections, especially fungal infections, such as histoplasmosis, blastomycosis, and coccidioidomycosis. Another increasingly common infection, tuberculosis, can also trigger it. Adrenal crisis is also found in patients with HIV infection, both as a direct cause of HIV or by superinfection by other organisms. Drugs may also cause adrenal dysfunction both by chronic immunosuppression by corticosteroids or by direct effect, such as with antifungal agents. Trauma, burns, and pancreatitis are all fairly common inciting events that trigger the cytokine cascade, leading to SIRS and distributive shock. Although this condition mimics the signs and symptoms of sepsis, this kind of distributive shock can be generated without an accompanying infection. Hypovolemic Shock Causes of hypovolemia that may lead to shock include loss of intravascular vol- ume through dehydration (from low fluid intake, diarrhea, bowel obstruction, sweating, or diabetes insipidus), diuresis (from diuretics or elevated blood glu- cose levels), capillary leak and third spacing (from burns, sepsis, pancreatitis, or surgical stress), hemorrhage (from trauma, gastrointestinal bleeding, fractures, vascular injuries, ruptured ovarian cysts, ectopic pregnancy, placental abruption, ruptured uterus, placenta previa), and anemia. Obstructive Shock Cardiac tamponade and restrictive pericarditis are characteristic of extracardiac obstructive shock. Pulmonary embolism caused by air, amniotic fluid, fat, or vas- cular clot can cause obstructive shock. Intrathoracic processes, such as pneu- mothorax, pulmonary hypertension, and diaphragmatic rupture, can all cause increased intrathoracic pressure and a decrease in forward flow of blood. The diagnosis of pulmonary processes may be made by chest x-ray films. Car- diac processes can be evaluated by rapid echocardiographic examination. Pul- monary embolism can be evaluated by ventilation/perfusion scans, echocardio- graphy, and pulmonary angiography. Cardiogenic Shock The most common cause of cardiogenic shock is an acute myocardial infarction (aMI). Myocardial infarction (MI) that affects cardiac valves can present with acute heart failure, as can MI of the left ventricular (LV) wall. Septal infarctions can lead to septal defects. Multiple infarctions can also affect already damaged myocardium, which may lead to rapid failure. Various cardiomyopathies (e.g., viral, alcoholic, infectious) may lead to cardiogenic shock. A post-MI arterial thrombus or aneurysms of the ventricular wall may also precipitate cardiac fail- ure. An MI that is well-tolerated initially but then extends (for hours to days) to involve a large degree of LV myocardium may be the sequence that most fre- quently leads to shock.

64 The Intensive Care Manual Severely reduced LV contractility induced by ischemia is the fundamental finding in cardiogenic shock. Extensive right ventricular (RV) infarction, which classically accompanies inferior-wall MI, can also present with hypotension. Cardiogenic shock is assessed by observing the patient’s mental status and skin color; by measuring urine output and blood pressure; and by evaluating for di- aphoresis, turgor, and tachycardia. Pulmonary congestion is evaluated by auscul- tation of the lungs for wet sounds (rales) and by the presence of S3 and S4 heart sounds. Measuring the levels of cardiac enzymes, myoglobins, and troponins is a rapid screening test and should be done along with evaluation of the ECG. Echocardiography and cardiac catheterization are the current gold standards: echocardiography rapidly shows both the location and extent of cardiac failure. In cardiogenic shock, forward blood flow is impaired by pump failure; and the typical picture is one of congestive heart failure (CHF), with increased fluid in the lung, pulmonary edema, and hepatic congestion. The hemodynamic picture is one of decreased cardiac output, elevated PCWP, elevated LV filling pressure, and in- creased SVR. The echocardiographic profile shows decreased cardiac function, wall-motion problems, decreased stroke volume, and decreased aortic diameter. MANAGEMENT AND THERAPY Patients suspected to be in shock should be managed in an ICU with close moni- toring and skilled nursing. Goals in the care of patients in different types of shock are the same, because the shock syndromes share many characteristics, regardless of their origin. The basic goal of shock therapy is the restoration of effective perfusion to vital organs and tissue before the onset of cellular injury. This basic therapy entails maintenance of appropriate cardiac function and mean arterial blood pressure. Endpoints as described in Table 3–4 can be used as a general guideline. Basic resuscitation should include rapid placement of a large-bore intra- venous line or a high-flow central line as a route for fluid resuscitation. Protec- tion of the airway through intubation and mechanical ventilation, if needed, can stabilize the patient. In unintubated patients, 2 to 15 L/min of high-flow oxygen is recommended to get oxygen saturation above 92%. To follow renal function, a Foley catheter should be placed early. Fluid Management Considerable quantities of fluid are often sequestered at a site of inflammation or lost because of fever, vomiting, or diarrhea. Because shock is greatly intensified when intravascular volume is depleted, fluid replacement is an important com- ponent of therapy. In a series of patients reported by Rackow, both cardiac output and survival were correlated with volume expansion.They demonstrated that increases in oxy-

3 / Shock 65 TABLE 3–4 General Goals for Support of Shock Patients Hemodynamic Support • MAP > 60–65 mm Hg • PCWP = 15–18 mm Hg • Cardiac index > 2.1 L/min per square meter of body surface area for cardiogenic and obstructive shock • Cardiac index > 4.0 L/min per square meter of body surface area for septic, traumatic, or hemorrhagic shock Optimization of Oxygen-Delivery • Hb level > 10 g/dL • Arterial oxygen saturation > 92% Reversal of Organ System Dysfunction • Maintain urine output > 0.5 mL/kg per hour ABBREVIATIONS: MAP, mean airway pressure; PCWP, pulmonary capillary wedge pressure; Hb, hemo- globin. gen consumption correlated with increases in cardiac output and oxygen delivery during fluid challenge of patients in septic shock. PAC monitoring and the newer technology of echocardiographic Doppler ul- trasonography have given the clinician more insight into the shock patient’s vol- ume status and can guide therapy to fixed endpoints. This ability to evaluate cardiac function may be an important advantage, because bacterial shock is often associated with impaired LV function. How much and how fast to give fluids is a matter of great debate. One tech- nique is to give the patient a fluid challenge, ranging from 5 to 20 mL/kg over a period of 10 minutes.11 The patient is then assessed for hemodynamic response (e.g., blood pressure, heart rate, urine output, mental status) and further fluid is administered based on the patient’s response. Alternatively, if the patient has central venous monitoring and the CVP pressure increases by more than 7 mm Hg above the initial pressure, the infusion is discontinued. If the pressure does not exceed the starting pressure by more than 3 mm Hg after 10 minutes of fluid infusion, or if it decreases by 3 mm Hg over a subsequent 10-minute rest period, a second aliquot of fluid is administered over 10 minutes and the “7 to 3 rule” is again applied.11 For plasma volume expansion, combinations of physiologic sodium solutions and plasma protein solutions are currently recommended. Albumin use has fallen over the past few years because of high costs and the lack of evidence-based outcome studies that it is superior to crystalloid solutions.12 Albumin may also induce activation of leukocytes at the endothelium, promoting the inflammatory cascade. Current data using different endpoints, such as mortality, length of ven- tilation, or renal function, do not show discernable differences between albumin and other colloids and crystalloids.

66 The Intensive Care Manual HYDROXYETHYL STARCHES The advantage of using hydroxyethyl starches (HES) is their prolonged duration in the circulation. The duration of HES as volume replacement is approximately 4 hours, with an intravascular half-life of 8 hours. HES are natural starches that have undergone hydroxylation and etherifica- tion, preventing hydrolysis by alpha-amylase. The main route of elimination of HES is renal, and use of HES should be avoided in patients with renal failure. A new formula of HES with physiologic solution (lactated Ringer’s solution) re- duces the chloride content and has been shown to have beneficial effects on organ function, with no coagulopathy determined.13 DEXTRAN AND GELATIN Use of dextran for volume replacement is vanishing in the United States because of the many side effects caused by these prepara- tions. Gelatins have not been approved for use in the United States because of the high incidence of allergic reactions, but they are used outside the United States. HYPERTONIC SALINE Hypertonic saline (2500 mOsm/L) increases plasma vol- ume by drawing fluid from the interstitial space.14 This “one-time” effect is sus- tained because of other mechanisms attributed to hypertonic saline, namely increased venous constriction and increased cardiac output. Hypertonic saline has added properties that may improve capillary flow and tissue oxygenation, and it now has wide use in neurosurgical and traumatic resuscitation. Hypertonic saline may increase the risk of arteriolar dilation and can theoretically produce or aggravate surgical bleeding. BLOOD REPLACEMENT Blood and component therapy is the only fluid that can correct anemia and coagulation problems. The optimal hematocrit for ICU patients remains to be determined. If blood transfusion were “risk-free,” the cur- rent intensive evaluation of transfusion practice would not be taking place. In the ICU, the risk of a blood transfusion–transmitted infection is not a major con- cern. What is of more consequence for the ICU patient is the evidence accumu- lating that blood transfusion has a profound negative effect on the immune system. While hemoglobin levels in the 7 to 10 mg/dL range are well tolerated in the “stable, non-stressed” patient, this range might not be optimal for the criti- cally ill patient. An intriguing new possibility is the use of erythopoietin in the critically ill population. Curwin at Dartmouth Medical Center has presented data that sug- gests that erythropoietin levels in the critically ill may be inappropriately low and that these patients may not be able to respond to endogenous erythropoietin (unpublished). CRYSTALLOIDS Physiologic isotonic crystalloid solutions are widely used for volume expansion. Although significant amounts are needed to restore the circu-

3 / Shock 67 lation, in the presence of hypovolemia the kinetics of distribution are altered and the 20% volume expansion may be increased significantly.15,16 In the presence of profound shock or massive volume loss, replacement with crystalloid solutions may require five times the volume lost, and the proportion may increase as more fluid is lost. Restoring the macrocirculation to normo- volemia does not necessarily restore the microcirculation or improve tissue oxy- genation. Edema must always be treated with care. The clinical significance of edema formation is unclear, but theoretically it can reduce oxygen transport. SUMMARY OF FLUID MANAGEMENT IN SHOCK Fluid replacement with crystalloid solutions is the mainstay of therapy, but this only treats the macrocir- culation. Better understanding of the endothelium as an organ has now led to an ongoing investigation into the different effects of fluids used for resuscitation.17 Perfection of these solutions to enable better microcirculation flow and modula- tion of the inflammatory response will lead to goal-directed therapy and out- come studies. Vasoactive Drugs The effectiveness of multiple vasoactive drugs for the treatment of shock has been studied, but no consensus has been reached. When using vasoactive drugs, how- ever, remember that the goal of this therapy is to increase perfusion to the tissues and not to artificially raise the blood pressure to an arbitrary goal. There is general agreement that volume control is the first-line treatment for shock. Fluid may stretch the left ventricle, leading to increased cardiac output, and refill the “tank.” This leads to increased blood pressure without vasoconstric- tion, which may actually reduce perfusion. The four most commonly used vasoactive agents in adult ICUs are dopamine, norepinephrine, phenylephrine, and dobutamine. The dosage ranges for these and other vasoactive drugs used in the ICU are listed in Table 3–5. In the US, the most commonly used vasopressor in patients with shock is dopamine, which is a naturally occurring precursor of norepinephrine that has different effects at different dosages. These effects, however, do overlap. TABLE 3–5 Commonly Administered Vasoactive Intravenous Drugs Dopamine, 1–20 µg/kg/min Norepinephrine, 0.05–2.0 µg/kg/min Dobutamine, 1–25 µg/kg/min Epinephrine, 0.05–2.0 µg/kg/min Phenylephrine, 2–10 µg/kg/min Isoproterenol, 1–8 µg/min Amrinone, 5–15 µg/kg per minute, after a 0.75 mg/kg bolus over 5 minutes Milrinone, 0.375–0.75 µg/kg per minute, after a 37.5–75 mg/kg bolus over 10 minutes

68 The Intensive Care Manual From a dosage of approximately 1 to 3 the direct dopaminergic effect of dopamine is seen. This results in vasodilatation of the renal and splanchnic circu- lations.18 The effects seen with the use of higher dosages of dopamine are mainly manifested through the endogenous release of norepinephrine. From 3 to 10 g/kg of body weight per minute, beta1 effects (increased inotropy and chronotropy) predominate. From 10 to 20 g/kg of body weight per minute, alpha1 activity (vasoconstriction) predominates. In patients with shock, lower doses are pre- ferred, with the goal of increasing forward flow and perfusion without inducing vasoconstriction, which may exacerbate ischemia. Norepinephrine is generally felt to be a second-line agent for shock in the United States; however, in Europe, it is a first-line agent. Norepinephrine has a mixture of beta- and alpha-agonist effects. The higher the dose, the more power- ful the vasoconstriction. Phenylephrine is a selective alpha-agonist and is used to increase blood pres- sure emergently, but it also decreases the microcirculation, so it cannot be used for long-term resuscitation. As it has no beta effect, it does not cause tach- yarrhythmias, and therefore may be useful in the cardiac patient who is prone to such problems. Dobutamine hydrochloride is a sympathomimetic agent with predominantly beta-adrenergic effects. This drug is not a vasopressor because it does not cause vasoconstriction, but it does increase forward flow. Its use in sepsis has not re- sulted in improved outcomes; its major use is to increase cardiac output in the failing heart. Dobutamine is not generally recommended in hypotensive patients because it results in reflex vasodilatation, which may manifest as hypotension, es- pecially in volume-depleted patients. In the treatment of cardiogenic shock with progressive perfusion failure, the ion-dilators, such as amrinone lactate, and dobutamine in combination with a vasopressor may allow cardiac function to improve. In other instances of CHF, there is a good indication for the administration of cardiac glycosides, preferably digoxin. In summary, the use of vasoactive drugs in shock is directed toward increasing perfusion. Vasopressor agents are used when blood pressure cannot be maintained with fluid alone. Continued monitoring for signs of perfusion inadequacy is neces- sary while using these agents. Finally, all vasopressor agents with alpha effects must be administered centrally, because if they were to infiltrate, they would cause local necrosis. If infiltration of any other vasopressor agents were to occur, the treatment is phentolamine (an alpha blocker) injected locally into the site of infiltration. Monoclonal Antibodies Multiple studies are ongoing in the use of monoclonal agents and receptor block- ers to modulate the immune system and the sepsis cascade. At this time bedside application of immunotherapeutic approached for the treatment of shock19 has not found clinical application.

3 / Shock 69 SUMMARY Our understanding of shock and SIRS response has evolved to one that is physio- logically based. Resuscitation is now based on close monitoring and hemody- namic support and replacement of intravascular volume. REFERENCES 1. Davies MD, Hagen PO. Systemic inflammatory response syndrome. Brit J Surg 1997;84:920–935. 2. Von Rueden TK, Dunham MC. Evaluation and management of oxygen delivery and consumption in multiple organ dysfunction syndrome in multiple organ dysfunction and failure, 2nd ed. In Secor VH, ed. Mosby Yearbook. St. Louis, MO: 1996:384–401. 3. Bone RC, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20:864–874. 4. Reddy PS, Curtiss EL, O’Toole JD, et al. Cardiac tamponade: Hemodynamic observa- tions in man. Circulation 1978;8:265–269. 5. Eisenberg MJ, Schiller NB. Bayes theorem and the echocardiographic diagnosis of cardiac tamponade. Am J Cardiol 1991;68:1242–1250. 6. Iberti TJ, Leibowitz AB, Papadakos PJ, et al. Low sensitivity of the anion gap as a screen to detect hyperlactemia in critically ill patients. Crit Care Med 1990; 18:275–277. 7. Rose S, Illerhaus M, Wiercinski A, et al. Altered calcium regulation and function of human neutrophils during multiple trauma. Shock 2000;13:92–99. 8. Muller-Berghaus G. Pathophysiologic and biochemical events in disseminated in- travascular coagulation: dysregulation of procoagulant and anticoagulant pathways. Seminar Thromb Hemost 1989;15:58–70. 9. Rasmussen HH, Ibel LS. Acute renal failure: Multivariate analysis of causes and risk factors. Am J Med 1982;733:211–218. 10. Cariou A, Mondi M, Luc-Marle J, et al. Noninvasive cardiac output monitoring by aortic blood flow determination: Evaluation of the Sometec Dynemo-3000 system. Crit Care Med 1988;12:2066–2072. 11. Packman MI, Rackow EC. Optimum left heart filling pressure during fluid resuscita- tion of patients with hypovolemic and septic shock. Crit Care Med 1983;11:165–169. 12. Cochran Injuries Group, Albumin Reviewers. Human albumin administration in critically ill patients: Systemic review of randomized controlled trials. Brit Med J 1998;317:235–40. 13. Treib J, Haass A, Pindur G, et al. All medium starches are not the same: Influence of the degree of hydroxyethyl substitution of hydroxyethyl starch on plasma volume, hemorrheologic conditions, and coagulation transfusion. Transfusion 1996;36:450–455. 14. Mattox KL, Maninagas PA, Moore EE, et al. Prehospital hypertonic saline/dextran infusion for post–traumatic hypotension: The USA multicenter trial. Ann Surg 1991; 213:482–491. 15. Drobin D. Volume kinetics of Ringer’s solution in hypovolemic volunteers. Anesthesi- ology 1999;90:81–91.

70 The Intensive Care Manual 16. Funk W, Balinger V. Microcirculatory perfusion using crystalloid or colloid in awake animals. Anesthesiology 1995;82:975–982. 17. Britt LD, Weireter LJ, Riblet JL, et al. Complex and challenging problems in trauma surgery. Surg Clin N Am 1996;76:645–660. 18. Lund N, DeAsla RJ, Guccione AL, et al. The effect of dopamine and dobutamine on skeletal muscle oxygenation in normoxemic rats. Cir Shock 1991;33:164–170. 19. Zeni F, Freeman B, Natanson C. Anti-inflammatory therapies to treat sepsis and sep- tic shock: A reassessment. Crit Care Med 1997;25:1095–1100.

CHAPTER 4 Approach to Mechanical Ventilation ANTHONY P. PIETROPAOLI INTRODUCTION Ventilator Settings Complications INVASIVE MECHANICAL Discontinuation of Noninvasive Mechanical VENTILATION Ventilation Indications CONCLUSION Objectives Modes Settings Mechanical Ventilation for Specific Conditions Complications Discontinuation of Mechanical Ventilation NONINVASIVE MECHANICAL VENTILATION Indications and Objectives Modes 71 Copyright 2001 The McGraw-Hill Companies. Click Here for Terms of Use.

72 The Intensive Care Manual INTRODUCTION Mechanical ventilation is defined as the use of a mechanical device to assist the respiratory muscles in the work of breathing and to improve gas exchange. In this chapter, mechanical ventilation is divided into two techniques: one requiring a tube in the trachea to deliver ventilation (invasive) and another applied with a mask (noninvasive). The indications, objectives, modes, settings, complications, and discontinuation strategies are reviewed for both invasive and noninvasive mechanical ventilation and some disease-specific strategies for invasive mechani- cal ventilation. INVASIVE MECHANICAL VENTILATION Indications Mechanical ventilation is indicated to support the patient with respiratory failure when adequate gas exchange cannot otherwise be maintained. As reviewed in chapter 1, there are two major categories of acute respiratory failure: hypoxemic (type 1) and hypercapneic (type 2). Patients with either of these often need me- chanical ventilation. Many patients present with a mixture of the two types of respiratory failure, and of course, these patients also respond to mechanical ven- tilation. Invasive mechanical ventilation is often chosen over noninvasive meth- ods when altered mental status or hemodynamic instability accompany acute respiratory failure. The timing of intubation and initiation of mechanical ventila- tion is a source of controversy, and the decision is often more a matter of art and experience than science. Tracheal intubation is indicated for situations other than provision of mechanical ventilation, such as to provide airway protection and relieve upper airway obstruction.1 Table 4–1 lists some commonly accepted indications for endotracheal intubation and mechanical ventilation. Objectives Mechanical ventilation is supportive and meant to reverse abnormalities in respi- ratory function, while specific therapies are used to treat the underlying cause of respiratory failure. The physiologic goals of mechanical ventilation are reversal of gas exchange abnormalities, alteration of pressure-volume relationships in the respiratory system, and reduction in the work of breathing.2 These physiologic goals are interrelated and attain specific clinical results, as shown in Figure 4–1. Other goals in specialized circumstances include allowing use of heavy sedation or neuromuscular blockade and stabilization of the chest wall when injury has disrupted its mechanical function.2

4 / Mechanical Ventilation 73 TABLE 4–1 Indications for Intubation and Invasive Mechanical Ventilation • Cardiac arrest • Respiratory arrest • Refractory hypoxemia (unresponsive to maximal supplemental oxygen administration and noninvasive ventilatory support) • Progressive respiratory acidosis (unresponsive to medical therapy, oxygen administra- tion, and noninvasive ventilatory support) • Symptoms of progressive respiratory fatigue (unresponsive to medical therapy, oxygen administration, and noninvasive ventilatory support) • Clinical signs of respiratory failure (unresponsive to medical therapy, oxygen adminis- tration, and noninvasive ventilatory support) • Tachypnea • Use of accessory muscles (e.g., sternocleidomastoid, scalene, intercostal, abdominal) • Paradoxical inward abdominal movement during inspiration • Progressive alteration of mental status • Inability to speak in full sentences • Airway protection (in a patient with an extremely impaired level of consciousness) • Relief of upper airway obstruction (often manifested by stridor on physical examina- tion) FIGURE 4–1 Objectives of mechanical ventilation. Interrelationship between physiologic ob- jectives of mechanical ventilation is shown. By accomplishing each of these physiologic objec- tives, specific clinical goals are met. (Adapted with permission from Slutsky AS. ACCP consensus conference: Mechanical ventilation. Chest 1993; 104(6):1833–1859.

74 The Intensive Care Manual Modes Mechanical ventilators were popularized during the polio epidemics of the 1950s. The initial ventilators were primarily negative pressure ventilators, or “iron lungs.” Later, positive pressure ventilators gained popularity and today are used almost exclusively. As ventilator technology has progressed, the ways of deliver- ing positive pressure mechanical ventilation have proliferated. In daily practice, however, four basic modes of positive pressure ventilation are most commonly used. These modes can be classified on the basis of how they are triggered to de- liver a breath, whether these breaths are targeted to a set volume or pressure, and how the ventilator cycles from inspiration to expiration (Table 4–2). CONTROLLED MECHANICAL VENTILATION Controlled mechanical ventila- tion (CMV) is included here only for the purposes of instruction. CMV, or vol- ume control (VC), was the first volume-targeted mode (Figure 4–2a). As its name suggests, it is a pure “control” mode; that is, the minute ventilation (VE,) is completely governed by the machine (VE = VT × respiratory rate). The physician sets the respiratory rate, tidal volume, inspiratory flow rate, ratio of inspiratory to expiratory time (I:E) fraction of inspired oxygen (FIO2), and positive end- expiratory pressure (PEEP). In VC, the patient is unable to trigger the ventilator to deliver additional breaths. This mode works well for patients who are unre- sponsive or heavily sedated, but not for conscious patients, whose respiratory ef- forts are not sensed by the ventilator, which leads to patient discomfort and increased work of breathing. As a result, this mode has largely been abandoned. ASSIST-CONTROL VENTILATION This mode is similar to VC mode except that the ventilator senses respiratory efforts by the patient (Figure 4–2b). As in VC, the physician sets a respiratory rate, tidal volume, flow rate, I:E, FIO2, and TABLE 4–2 Basic Modes of Mechanical Ventilation Mode Trigger Target Cycle Volume controla Ventilator VT Time and VT Assist-controla Ventilator ± patient VT Time and VT SIMVa Ventilator ± patient VT/VI (SIMV Time and VT/VT breaths only) (SIMV breaths only) Pressure-controlb Ventilator ± patient Inspiratory pressure Time Pressure-supportc Patient Inspiratory pressure Flow ABBREVIATIONS: SIMV, synchronized intermittent mandatory ventilation; VT, tidal volume; ± = with or without. aAll volume-targeted modes cycle from inspiration to expiration at the end of inspiratory time, which corresponds to the instant that the VT is reached. The target VT is achieved by setting a fixed inspira- tory flow for a fixed inspiratory time interval. bIn pressure-control mode, the desired pressure is achieved almost immediately after the onset of in- spiration. Target pressure is maintained for the duration of set inspiratory time. cIn pressure-support mode, the pressure target is maintained until inspiratory flow falls to about 20% of peak flow. Inspiratory time varies from breath to breath.

4 / Mechanical Ventilation 75 a. b. FIGURE 4–2 Airway opening pressure (PaO), lung volume (V), and inspiratory (I), and ex- piratory (E) flow rate (V) versus time during mechanical ventilation. a. Volume control (VC), also known as controlled mechanical ventilation (CMV). During both breaths shown, defined tidal volume (VT) and inspiratory flow rate are delivered, result- ing in PaO2 shown. In this mode, ventilator does not detect patient efforts. A reduction in air- way pressure from patient effort (arrow) does not result in significant VT or inspiratory flow. b. Assist-control (AC) ventilation. Notice that ventilator senses decrease in airway pressure induced by patient effort (indicated by arrow) and delivers same VT and flow in response.

76 The Intensive Care Manual c. d. FIGURE 4–2 (continued) c. Synchronized intermittent mandatory ventilation (SIMV). First breath is ventilator- delivered in absence of patient effort. Next, patient effort causes decrease in PaO during syn- chronization period (boxes), so fully supported breath is delivered. Next effort occurs outside of synchronization period, and patient breathes spontaneously. Resulting volume and pressure are completely patient-generated. Last breath is identical to first, delivered according to set respiratory rate. End of synchronization period coincides with onset of the back-up SIMV breath. d. Pressure-control (PC) ventilation. Airway pressure is set, and VT and flow rate that result are variable and depend on inspiratory time, airway resistance, respiratory system compli- ance, and patient effort. In example shown, patient is relaxed. First breath is delivered auto- matically by ventilator, based on fixed back-up respiratory rate. Second breath is delivered early, when patient lowers airway pressure and triggers ventilator (arrow).

4 / Mechanical Ventilation 77 e. FIGURE 4–2 (continued) e. Pressure-support (PS) ventilation. Inspiratory pressure is fixed in this mode, as in pressure- control mode. However, this mode is flow-cycled instead of time-cycled. Inspiratory pressure ceases when inspiratory flow rate decreases to about 20% of its peak. VT and flow are deter- mined by inspiratory pressure, airway resistance, respiratory system compliance, and patient effort. First breath shows moderate inspiratory effort. In second example, patient makes a pro- longed inspiratory effort, resulting in more prolonged delivery of inspiratory pressure and a larger VT. Third example shows rapid deep breath, resulting in very high peak inspiratory flow rate but short duration of inspiratory pressure. The resulting VT is midway between other two examples. (Modified with permission, from Schmidt GA, Hall JB. Management of the ventilated patient. In Hall JB, Schmidt GA, Wood LDH, eds. Principles of critical care, 2nd ed. New York: McGraw-Hill, 1998:517–535.) PEEP. Breaths are delivered automatically, regardless of patient effort (“con- trol”). In assist-control (AC) mode, however, the ventilator detects patient effort and responds by delivering a breath identical to the controlled one (“assist”). The patient can therefore breathe faster than the back-up control rate, but all breaths have the same tidal volume, flow rate, and inspiratory time. So AC mode allows better synchrony between patient and ventilator than VC mode, while still pro- viding a baseline minute ventilation. A more descriptive and accurate name for this mode is “volume-targeted assist-control ventilation.” However, the term “AC” is well entrenched and likely will not be replaced by this more cumbersome name. Like all modes of mechanical ventilation, AC has disadvantages. If the back-up respiratory rate is set too far below the patient’s spontaneous rate, exhalation time progressively decreases, since inspiratory time is fixed by the back-up respi-

78 The Intensive Care Manual ratory rate and flow rate. In the extreme, this may result in inadequate time for exhalation (Figure 4–3). As a result, lung volume remains above functional resid- ual capacity (FRC) when the next breath is delivered, a process called dynamic hyperinflation.2 This increased lung volume is associated with elevation in the alveolar pressure at end-exhalation, or “auto-PEEP” (Figure 4–3). The adverse consequences of these events are discussed later. Another problem occurs when patients with high minute ventilation requirements make persistent inspiratory efforts while a breath is being delivered. If this effort is strong enough, the patient FIGURE 4–3 Dynamic hyperinflation and auto-PEEP (positive end-expiratory pressure) re- sult from inadequate exhalation time. Simplified schematic shows two lung units, consisting of alveolus and airway, both at end exhalation. In a, there is adequate time for complete ex- halation to resting lung volume, or functional residual capacity (FRC). The alveolar pressure is zero, or equal to level of externally applied PEEP. In b, there is inadequate time for exhala- tion. This occurs when exhalation time is too short and/ or time required to exhale to FRC is pathologically prolonged. Former occurs during mechanical ventilation when inspiratory time is too long or respiratory rate is too high; latter occurs in obstructive lung diseases, like chronic obstructive pulmonary disease (COPD) and asthma. In either case, lung volume remains above FRC at end exhalation (dynamic hyperinflation), resulting in abnormally elevated PA (auto-PEEP).

4 / Mechanical Ventilation 79 may trigger the ventilator again, a phenomenon known as “breath stacking.” This can cause wide swings in airway pressure and increase the risk of barotrauma or ventilator-associated lung injury. Finally, in volume-targeted modes, the inspira- tory flow rate is fixed. Many acutely ill patients strive for high inspiratory flow rates. If ventilator delivered air flow is below patient demand, the work of breath- ing increases as the patient makes futile efforts to augment inspiratory flow. SYNCHRONIZED INTERMITTENT MANDATORY VENTILATION Like AC mode, synchronized intermittent mandatory ventilation (SIMV) is also a volume-targeted mode and provides a guaranteed VE (Figure 4–2c). For the mandatory breaths, tidal volume and respiratory rate are chosen, guaranteeing a baseline minute ventilation. The practitioner also sets FIO2, PEEP, and flow rate. As in AC mode, the patient can make inspiratory efforts between the mandatory breaths. If a sufficient effort occurs shortly before the mandatory breath is deliv- ered (a time interval known as the “synchronization period”), a breath identical to the mandatory breath is delivered. If a patient effort occurs outside this syn- chronization period, the airway pressure, flow rate, and tidal volume are purely patient-generated, and no assistance is provided by the ventilator. While this re- duces the likelihood of air-trapping and breath-stacking, it also can increase the work of breathing. Interestingly, if the mandatory respiratory rate is less than ap- proximately 80% of the patient’s actual rate, the high level of work expended during the spontaneous breaths will also be expended during the mandatory breaths.3 This occurs because the respiratory center in the brain has a lag time and is unable to alter its output on a breath-to-breath basis. So if high neurologic output is required for a significant percentage of breaths, that same output will be given for all of the breaths, including those that are delivered by the ventilator. Therefore, attempting to “exercise” the respiratory muscles by setting the SIMV rate at half of the patient’s spontaneous rate is counterproductive, because it sim- ply increases the work of breathing and results in respiratory muscle fatigue and weaning failure. To prevent excessive work while still allowing the patient to breath above the SIMV rate, this mode is often combined with pressure-support ventilation, discussed later. PRESSURE-CONTROL VENTILATION A more accurate name for pressure- control ventilation (PCV) mode is “pressure targeted assist-control ventilation” (Figure 4–2d). The mode is similar to the assist-control mode described above, except that a defined inspiratory pressure (IP) is set, instead of a tidal volume (Figure 4–2d). This allows absolute control over peak pressure delivered by the ventilator, which can have advantages in certain types of lung disease. Other de- fined settings are similar to assist-control: respiratory rate, I:E ratio, FIO2, PEEP, and trigger sensitivity. When the ventilator detects patient effort, it delivers a breath identical to the backup-controlled breaths, allowing the patient to breathe faster than the back-up rate. Tidal volume is determined by IP, inspiratory time, airway resistance, respiratory system compliance (∆∆VP), and patient effort. The de-

80 The Intensive Care Manual livered volume is predictable if sufficient time is given to allow equalization be- tween the delivered inspiratory pressure and alveolar pressure.4 If inspiratory time is too short or airway resistance is too high, this equilibration does not occur, resulting in a tidal volume lower than predicted and a decrease in minute ventilation. In response, the patient increases respiratory rate. Paradoxically, the increase in respiratory rate causes a decrease in minute ventilation because, as res- piratory rate increases, expiratory time also decreases. The result is inadequate time for complete exhalation, dynamic hyperinflation, and auto-PEEP. The re- sulting decrease in respiratory system compliance reduces the tidal volume at- tained for the given IP. This is one of the major disadvantages of PCV, and is most often seen in the setting of obstructive lung disease. Inspiratory flow rate is not fixed in PCV. It varies with IP, inspiratory time, respiratory mechanics, and patient effort. This can be advantageous, because flow rate increases with patient effort, unlike the volume-targeted modes, in which flow rate is fixed. As a result, patients with high minute-ventilation requirements may feel more comfortable on PCV, because they can regulate and increase flow as needed. This variable flow rate has another potential advantage: the flow pat- tern changes as respiratory system compliance decreases during lung inflation. Thus, flow is high early in inspiration when the system is very compliant and de- creases as inflation proceeds and compliance decreases. The result is a lower peak airway pressure and a flow pattern that more closely mimics normal physiology. Whether this leads to any improvements in clinical outcome is unclear. PRESSURE-SUPPORT VENTILATION The unique feature of pressure-support ventilation (PSV) is that it is flow-cycled instead of time-cycled (Figure 4–2e). So IP ceases when the flow rate drops to about 20% of peak flow rate, and passive exhalation occurs. The practitioner sets pressure-support level, FIO2 and PEEP. Respiratory rate, inspiratory flow rate, tidal volume, and I:E ratio are determined by the patient’s effort and respiratory system mechanics (resistance and compli- ance). PSV is an “apnea mode,” that is, there is no back-up mandatory respira- tory rate, so it can only be used for patients with adequate respiratory drive. PSV is often combined with SIMV. This reduces the work of breathing in comparison to SIMV alone and provides a back-up mandatory minute ventila- tion not available with PSV alone. ALTERNATIVE MODES The number of available modes of ventilation has in- creased rapidly. These include high-frequency ventilation, airway pressure- release ventilation, proportional-assist ventilation, and servo-controlled pressure support modes. A review of these modes is beyond the scope of this chapter, and the reader is referred to in-depth discussions of mechanical ventilation5 and a re- cent review article.6

4 / Mechanical Ventilation 81 Settings The parameters that need to be set vary, depending on the mode of ventilation used, as demonstrated in Table 4–3. Initial values for the different ventilator set- tings are shown in Table 4–4.2 RESPIRATORY RATE There is a wide range of mandatory ventilator-delivered res- piratory rates that can be used. The number varies and is dependent on the minute ventilation goal, which varies with individual patients and clinical circumstances. In general, the range for respiratory rate is between 4/min and 20/min and falls be- tween 8/min and 12/min in most stable patients.2 In adult respiratory distress syn- drome (ARDS), the use of low tidal volumes sometimes necessitates respiratory rates up to 35/min to maintain adequate minute ventilation.7 TIDAL VOLUME Evidence is accumulating that tidal volumes should be lower than traditionally recommended, especially in acute respiratory distress syn- drome.7,8,9 When setting tidal volume in volume-targeted modes, a rough estimate for patients with lung disease is 5 to 8 mL/kg of ideal body weight. In patients with normal lungs who are intubated for other reasons, slightly higher tidal volumes can be considered: up to 12 mL/kg of ideal body weight. Tidal volume should be ad- justed to maintain a plateau pressure of less than 35 cm H2O. The plateau pressure is determined by performing an inspiratory-hold maneuver (Figure 4–4a), which approximates end-inspiratory alveolar pressure in a relaxed patient. Elevation in the plateau pressure may not always increase the risk of baro- trauma. This risk rises with transalveolar pressure, which is the alveolar pressure minus the pleural pressure. In patients with chest-wall edema, abdominal disten- tion, or ascites, compliance of the chest wall is reduced. As a result, pleural pres- sure rises during lung inflation and the rise in transalveolar pressure is lower than TABLE 4–3 Required Settings for Different Ventilator Modes Setting VC AC SIMV PC PS Rate    VT   IP  TS    Flow rate   I:E    FIO2    PEEP   ABBREVIATIONS: VC, volume control; AC, assist-controlled; SIMV, synchronized intermittent manda- tory ventilation; PC, pressure-control; PS, pressure-support; VT, tidal volume; IP, inspiratory pres- sure; TS, trigger sensitivity; I:E, ratio of inspiratory to expiratory time; FIO2, fraction of inspired oxygen; PEEP, positive end-expiratory pressure.

82 The Intensive Care Manual TABLE 4–4 Suggestions for Initial Ventilator Settings Parameter Usual Range Adjust to Maintain Rate (breaths/min) 4–20 breaths/min Patient comfort, pH > 7.25, avoid auto-PEEP VT Lung disease: 5–8 mL/kg Plateau pressure ≤ 35 cm H2O Normal: 8–12 mL/kg IP 10–30 cm H2O Plateau pressure ≤ 35 cm H2O FIO2 0.3–1.0% O2 sat ≤ 90%, FIO2 ≤ 0.6% PEEP 3–20 cm H2O Plateau pressure ≤ 35 cm H2O, O2 sat ≥ 90% TS Pressure: −1–2 cm H2O Patient triggering ventilator Flow −1–3 L/min effectively Flow rate 40–100 L/min Patient comfort; avoid auto-PEEP I:E 1:1.5 to 1:3 Patient comfort; avoid auto-PEEP ABBREVIATIONS: PEEP, positive end = expiratory pressure; VT, tidal volume; IP, inspiratory pressure; FIO2, fraction of inspired oxygen; TS, trigger sensitivity; ; I:E, ratio of inspiratory to expiratory time. occurs with normal chest compliance. In such circumstances, the tidal volume ranges previously discussed should be used. INSPIRATORY PRESSURE In PCV and PSV, the IP is generally set to keep the plateau pressure at or below 35 cm H2O. The resulting tidal volume should be kept in the suggested ranges. FRACTION OF INSPIRED OXYGEN In most cases, FIO2 should be 100% when the patient is first intubated and placed on mechanical ventilation. Once proper tube placement is assured and the patient has stabilized, FIO2 should be progres- sively reduced to the lowest concentration that maintains adequate oxygen satu- ration of hemoglobin, because high concentrations of oxygen produce pulmonary toxicity. Maintaining oxygen saturation of 90% or more is the usual goal. Occa- sionally, this goal is superseded by the need to protect the lung from excessive tidal volumes, pressures, or oxygen concentrations. In these circumstances, the target may be lowered to 85%, while optimizing the other factors involved in oxygen delivery (see chapter 1). POSITIVE END-EXPIRATORY PRESSURE PEEP, as its name implies, maintains a set level of positive airway pressure during the expiratory phase of respiration. It dif- fers from continuous positive airway pressure (CPAP) in that it is only applied dur- ing expiration, whereas CPAP is applied throughout the entire respiratory cycle. The use of PEEP during mechanical ventilation has several potential benefits. In acute hypoxemic respiratory failure (type 1), PEEP increases mean alveolar pres- sure, promotes re-expansion of atelectatic areas, and may force fluid from the alve- olar spaces into the interstitium. This allows previously closed or flooded alveoli to participate in gas exchange. In cardiogenic pulmonary edema, PEEP can reduce left ventricular preload and afterload, improving cardiac performance.

4 / Mechanical Ventilation 83 a. b. FIGURE 4–4 Determining plateau pressure and auto-PEEP. a. Method for determining plateau pressure. Graphs of airway pressure, volume, and flow ver- sus time are shown during volume-targeted ventilation. An inspiratory pause is performed in relaxed patient by occluding airway at end-inspiration (thick arrow). Pressure drops from peak to plateau as flow stops and end-inspiratory volume is maintained. When airway occlu- sion is released, expiratory flow occurs and lung volume returns to FRC. b. Method for estimating auto-PEEP. An expiratory pause is performed in a relaxed patient by occluding airway at end-expiration (thick arrow). Measured pressure rises as flow stops and PA equilibrates with airway pressure. The next breath from ventilator causes flow to re- sume, and airway pressure and lung volume rise. (Modified with permission from Aldrich TK, Prezant DJ. Indications for mechanical ventilation. In Tobin MJ, ed. Principles and practice of mechanical ventilation. New York: McGraw-Hill, 1994:155–189.)